Hepatocellular carcinoma (HCC) is the third cause of cancer-related death worldwide with the most rapidly rising cancer incidence in the US and Western Europe. The disease is a major public health problem, for which there are no molecular diagnostics and limited effective treatment options. The hypothesis of the project is that knowledge of the gene expression profiles and the somatic genetic alterations of HCC will enable the identification of tumor suppressor genes or oncogenes involved in the pathogenesis of the disease, and will provide markers of early detection and new treatment targets. The identification of gene signatures predictive of clinical outcomes will enable a molecular classification of HCC. For that purpose we will test 230 samples [including 20 dysplastic nodules and 100 paired HCV- related HCC/non-tumoral cirrhotic liver]. We already have collected 214 samples from patients in 3 referral HCC Units: Mount Sinai, New York; Hospital Clinic Barcelona (Collaborator #2); National Cancer Institute, Milan (Collaborator #3). Specific aims: Aim #1: Characterize the gene expression profiles of dysplastic lesions and hepatocellular carcinoma in HCV infected patients. 1a. Determine the genes and pathways involved in hepatocarcinogenesis, by genome- wide gene expression analysis. By using genome-wide microarray technology profiling of 38,500 genes, (Affymetrix, Genechip Human Genome-U133 Plus 2.0), we have identified molecular signatures that classify dysplastic nodules, early tumors and advanced cancers in a preliminary set of 73 samples. We will expand the analysis to 230 samples. 1b. Identify a molecular signature predictive of HCC recurrence and survival. A gene signature able to discriminate good and poor outcomes will be analyzed in 85 patients undergoing resection for HCC in which complete clinical data has been recorded. Multivariate approaches will ensure a proper integration of the clinical and molecular data. Aim #2: To define the extent of structural alterations in the HCC genome using high-density single nucleotide polymorphisms arrays interrogating 500,000 genomic loci. To integrate emerging genetic maps with paired expression data to identify candidate tumor suppressors and oncogenes. We will apply high-density 500K SNP arrays to the analysis of structural somatic genetic alterations in HCC, in collaboration with the Dr. William Sellers/Matthew Meyerson's group in Dana-Farber Institute, Harvard (Collaborator #1). By using integrative genomic analysis of the molecular profiling and genetic disturbances we'll identify novel tumor suppressors or oncogenes, which may constitute new targets for chemopreventive and therapeutic strategies. Aim #3: Identify molecular biomarkers for the diagnosis of early HCC. #3a: Identification of biomarkers. We have identified by Real-Time RT-PCR a gene signature of 3 genes (Glypican-3, Survivin, LYVE-1) able to discriminate between dysplastic lesions from early cancer with an accuracy of 94%. By microarray technology, we have preliminary data with a set of 93 potential candidate genes for the identification of new biomarkers. #3b: To validate the molecular signature for the diagnosis of early HCC in cirrhotic patients undergoing screening. The gene- signature identified will be validated in the clinical practice in a testing set of 80 samples from nodules between 0.5-2cm obtained from a prospective cohort of HCV-cirrhotic patients (Collaborator #2, University of Barcelona). [unreadable] [unreadable] [unreadable]